Developer Builds Free Local Voice-to-Text Pipeline Using Node.js and Whisper
A developer created a fully local audio transcription pipeline after finding cloud-based APIs from Google and AWS would cost $50–$100 per month. The system uses FFmpeg to normalize audio, OpenAI's Whisper (via whisper.cpp) for transcription, and Node.js to tie the stages together. Output is saved as SRT subtitle files and stored in a SQLite database using the better-sqlite3 library. The pipeline runs entirely on a local GPU, making it a cost-free alternative to commercial transcription services. The developer shared the full Node.js source code, covering audio normalization, Whisper CLI integration, and database storage.
This is an AI-generated summary. ShortSingh links to the original source for the complete article.
Discussion (0)
Log in to join the discussion and vote.
Log in